https://aiweekly.co/alerts/picbreeder-replayed-with-vlms-to-probe-ai-open-endedness AI Weekly en 2026-08-20T06:12:01+00:00 Picbreeder replayed with VLMs to probe AI open-endedness https://aiweekly.co/alerts/claude-opus-48-golds-iol-ai-linguistics-reasoning-benchmark AI Weekly en 2026-08-20T06:12:01+00:00 Claude Opus 4.8 golds IOL-AI linguistics reasoning benchmark https://aiweekly.co/alerts/amazon-bedrock-adds-xais-grok-46-with-500k-context-window AI Weekly en 2026-08-20T05:02:29+00:00 Amazon Bedrock adds xAI's Grok 4.6 with 500K context window https://aiweekly.co/alerts/google-cloud-to-hire-hundreds-of-forward-deployed-ai-engineers AI Weekly en 2026-08-20T05:02:29+00:00 Google Cloud to hire hundreds of forward-deployed AI engineers https://aiweekly.co/alerts/watermarks-degrade-clinical-llm-reasoning-multi-model-study-finds AI Weekly en 2026-08-20T03:54:21+00:00 Watermarks Degrade Clinical LLM Reasoning, Multi-Model Study Finds https://aiweekly.co/alerts/study-agent-skills-stabilize-procedures-657-vs-45-knowledge AI Weekly en 2026-08-20T03:54:21+00:00 Study: agent skills stabilize procedures, 65.7% vs 4.5% knowledge https://aiweekly.co/alerts/v-rae-tops-video-vaes-on-kinetics-600-converges-6x-faster AI Weekly en 2026-08-20T03:54:21+00:00 V-RAE tops video VAEs on Kinetics-600, converges 6x faster https://aiweekly.co/alerts/meta-smart-glasses-recording-light-easy-to-miss-guardian-says AI Weekly en 2026-08-20T03:12:02+00:00 Meta smart glasses' recording light 'easy to miss', Guardian says https://aiweekly.co/alerts/multi-agent-ai-coders-reach-for-hidden-grading-files-in-study AI Weekly en 2026-08-20T03:12:02+00:00 Multi-agent AI coders reach for hidden grading files in study https://aiweekly.co/alerts/cursor-gives-cloud-agents-subscriptions-goal-and-subagent-vms AI Weekly en 2026-08-20T01:58:49+00:00 Cursor gives cloud agents subscriptions, /goal and subagent VMs https://aiweekly.co/alerts/bjar-tells-jury-metas-teen-safeguards-were-designed-to-fail AI Weekly en 2026-08-20T00:31:40+00:00 Béjar tells jury Meta's teen safeguards were 'designed to fail' https://aiweekly.co/alerts/xiaomis-17-meter-humanoid-set-for-public-debut-in-beijing AI Weekly en 2026-08-20T00:31:40+00:00 Xiaomi's 1.7-Meter Humanoid Set for Public Debut in Beijing https://aiweekly.co/alerts/physics-of-agents-paper-ai-groups-drift-right-on-politics AI Weekly en 2026-08-20T00:12:02+00:00 Physics of Agents paper: AI groups drift right on politics https://aiweekly.co/alerts/doctorow-argues-doge-is-vibe-governance-by-ai-chatbot AI Weekly en 2026-08-20T00:12:02+00:00 Doctorow argues DOGE is 'vibe governance' by AI chatbot https://aiweekly.co/alerts/doctorow-chatbots-deliver-delusional-reinforcement-as-a-service AI Weekly en 2026-08-20T00:12:02+00:00 Doctorow: chatbots deliver 'delusional reinforcement as a service' https://aiweekly.co/alerts/marvell-hands-google-122b-warrant-in-custom-chip-deal AI Weekly en 2026-08-19T21:31:52+00:00 Marvell hands Google $12.2B warrant in custom-chip deal https://aiweekly.co/alerts/fractile-in-talks-for-600m-at-65b-on-anthropic-chip-deal AI Weekly en 2026-08-19T21:31:52+00:00 Fractile in talks for $600M at $6.5B on Anthropic chip deal https://aiweekly.co/alerts/flocks-os-investigate-lets-police-search-by-movement-pattern AI Weekly en 2026-08-19T21:31:52+00:00 Flock's OS Investigate lets police search by movement pattern https://aiweekly.co/alerts/amazon-drops-1999-alexa-fee-makes-it-free-on-fire-tv AI Weekly en 2026-08-19T21:31:52+00:00 Amazon drops $19.99 Alexa+ fee, makes it free on Fire TV https://aiweekly.co/alerts/flocks-os-investigate-lets-police-track-drivers-by-patterns AI Weekly en 2026-08-19T21:12:01+00:00 Flock's OS Investigate lets police track drivers by patterns https://aiweekly.co/alerts/berkeley-law-bans-ai-drafting-exam-use-starting-summer-2026 AI Weekly en 2026-08-19T21:12:01+00:00 Berkeley Law bans AI drafting, exam use starting Summer 2026 https://aiweekly.co/alerts/vauhini-vara-tests-whether-readers-can-tell-ai-from-her-fiction AI Weekly en 2026-08-19T21:12:01+00:00 Vauhini Vara tests whether readers can tell AI from her fiction https://aiweekly.co/alerts/truefoundry-open-sources-trueforge-a-claude-agents-rival AI Weekly en 2026-08-19T21:12:01+00:00 TrueFoundry open-sources TrueForge, a Claude Agents rival https://aiweekly.co/alerts/openai-slows-astra-training-after-agent-breaches-hugging-face AI Weekly en 2026-08-19T21:12:01+00:00 OpenAI Slows Astra Training After Agent Breaches Hugging Face https://aiweekly.co/alerts/taos-icm-2026-essay-sets-ground-rules-for-ai-in-mathematics AI Weekly en 2026-08-19T21:12:01+00:00 Tao's ICM 2026 essay sets ground rules for AI in mathematics https://aiweekly.co/alerts/unsloth-ships-1-bit-to-bf16-gguf-quants-of-qwen38-27b AI Weekly en 2026-08-19T21:12:01+00:00 Unsloth ships 1-bit to BF16 GGUF quants of Qwen3.8-27B https://aiweekly.co/alerts/deepreinforce-releases-ornith-15-in-397b-35b-and-9b-sizes AI Weekly en 2026-08-19T21:12:01+00:00 DeepReinforce releases Ornith-1.5 in 397B, 35B and 9B sizes https://aiweekly.co/alerts/bioethics-paper-argues-right-to-refuse-ai-chatbots-on-campus AI Weekly en 2026-08-19T21:12:01+00:00 Bioethics paper argues right to refuse AI chatbots on campus https://aiweekly.co/alerts/pew-52-of-americans-now-more-concerned-than-excited-about-ai AI Weekly en 2026-08-19T21:12:01+00:00 Pew: 52% of Americans now more concerned than excited about AI https://aiweekly.co/alerts/radboud-faculty-doc-sets-do-not-use-default-for-genai AI Weekly en 2026-08-19T21:12:01+00:00 Radboud faculty doc sets 'do not use' default for GenAI https://aiweekly.co/alerts/hugging-faces-teen-chatbot-adelina-wins-over-a-skeptic AI Weekly en 2026-08-19T21:12:01+00:00 Hugging Face's teen chatbot Adelina wins over a skeptic https://aiweekly.co/alerts/fda-weighs-clinician-style-tests-for-genai-medical-devices AI Weekly en 2026-08-19T19:58:37+00:00 FDA weighs clinician-style tests for genAI medical devices https://aiweekly.co/alerts/teen-boys-use-meta-ray-bans-to-film-girls-in-us-schools AI Weekly en 2026-08-19T19:58:37+00:00 Teen Boys Use Meta Ray-Bans to Film Girls in US Schools https://aiweekly.co/alerts/rivian-spinout-also-raises-150m-series-d-led-by-prysm AI Weekly en 2026-08-19T18:38:04+00:00 Rivian spinout Also raises $150M Series D led by Prysm https://aiweekly.co/alerts/meta-ships-mac-app-for-meta-ai-with-ads-and-workspace-hooks AI Weekly en 2026-08-19T18:38:04+00:00 Meta Ships Mac App for Meta AI With Ads and Workspace Hooks https://aiweekly.co/alerts/openai-pilots-private-safety-processing-for-paid-api-tier AI Weekly en 2026-08-19T18:38:04+00:00 OpenAI pilots Private Safety Processing for paid API tier https://aiweekly.co/alerts/big-tech-opens-data-center-doors-as-130b-in-ai-projects-stall AI Weekly en 2026-08-19T09:32:56+00:00 Big Tech opens data-center doors as $130B in AI projects stall https://aiweekly.co/alerts/llm-verifiers-grow-more-lenient-after-audit-repair-in-context AI Weekly en 2026-08-19T09:12:01+00:00 LLM verifiers grow more lenient after audit-repair in context https://aiweekly.co/alerts/residual-signatures-reveal-llm-lineage-from-weights-alone AI Weekly en 2026-08-19T09:12:01+00:00 Residual signatures reveal LLM lineage from weights alone https://aiweekly.co/alerts/stanfords-zou-puts-agentic-ai-to-work-across-biomed-research AI Weekly en 2026-08-19T09:12:01+00:00 Stanford's Zou puts agentic AI to work across biomed research https://aiweekly.co/alerts/cerebras-cs-4-posts-4400-tokenssec-up-to-30x-vs-gpus AI Weekly en 2026-08-19T09:12:01+00:00 Cerebras CS-4 posts 4,400 tokens/sec, up to 30x vs GPUs https://aiweekly.co/alerts/ai-observatory-vendor-usage-reports-filter-half-of-chats AI Weekly en 2026-08-19T06:12:01+00:00 AI Observatory: vendor usage reports filter half of chats https://aiweekly.co/alerts/microverse-probes-identity-drift-in-multi-agent-llm-simulations AI Weekly en 2026-08-19T06:12:01+00:00 MicroVerse probes identity drift in multi-agent LLM simulations https://aiweekly.co/alerts/statem-953-on-terminal-bench-or-a-15-frontier-run-via-harness AI Weekly en 2026-08-19T03:56:04+00:00 StateM: 95.3% on Terminal-Bench or a $15 Frontier Run via Harness https://aiweekly.co/alerts/freetoken-paper-753b-glm-52-moe-on-one-workstation-gpu AI Weekly en 2026-08-19T03:56:04+00:00 FreeToken paper: 753B GLM-5.2 MoE on one workstation GPU https://aiweekly.co/alerts/harmprofile-finds-harm-scales-with-capability-across-23-llms AI Weekly en 2026-08-19T03:56:04+00:00 HarmProfile finds harm scales with capability across 23 LLMs https://aiweekly.co/alerts/abra-study-compute-optimal-diffusion-needs-200-tokensparam AI Weekly en 2026-08-19T03:56:04+00:00 Abra study: compute-optimal diffusion needs 200 tokens/param https://aiweekly.co/alerts/asi-bench-halves-18-frontier-agents-scores-without-guidance AI Weekly en 2026-08-19T03:56:04+00:00 ASI-Bench halves 18 frontier agents' scores without guidance https://aiweekly.co/alerts/ai-observatory-anthropic-method-filters-48-of-real-ai-chats AI Weekly en 2026-08-19T03:30:05+00:00 AI Observatory: Anthropic method filters 48% of real AI chats https://aiweekly.co/alerts/anthropic-releases-dataset-behind-claudes-1415-binder-run AI Weekly en 2026-08-19T03:12:01+00:00 Anthropic releases dataset behind Claude's 14/15 binder run https://aiweekly.co/alerts/language-models-dont-reliably-encode-the-current-year AI Weekly en 2026-08-19T03:12:01+00:00 Language models don't reliably encode the current year https://aiweekly.co/alerts/cleanslate-picking-what-to-forget-breaks-llm-unlearning AI Weekly en 2026-08-19T03:12:01+00:00 CleanSlate: picking what to forget breaks LLM unlearning https://aiweekly.co/alerts/semianalysis-core-research-eyes-100m-year-on-buy-side-demand AI Weekly en 2026-08-19T03:12:01+00:00 SemiAnalysis Core Research eyes $100M year on buy-side demand https://aiweekly.co/alerts/llm-watermarks-degrade-medical-reasoning-eth-paper-finds AI Weekly en 2026-08-19T00:12:01+00:00 LLM watermarks degrade medical reasoning, ETH paper finds https://aiweekly.co/alerts/stellenbosch-dean-flagged-by-ai-detector-rethinks-authorship AI Weekly en 2026-08-19T00:12:01+00:00 Stellenbosch dean flagged by AI detector rethinks authorship https://aiweekly.co/alerts/lens-drops-persistent-indexes-for-query-time-evidence-search AI Weekly en 2026-08-19T00:12:01+00:00 LENS drops persistent indexes for query-time evidence search https://aiweekly.co/issues/what-ai-models-are-actually-coming-in-the-next-six-months AI Weekly en 2026-08-20T12:00:00+00:00 What AI models are actually coming in the next six months?